Page 59 - Banking Finance May 2024

ARTICLE

nonprofit organizations, to participate in the social bond Conclusion
market. Diversifying the pool of issuers and projects can Social bonds play a crucial role in promoting sustainable
expand the universe of investable opportunities and
finance, driving positive social impact, and aligning financial
address a wider range of social challenges.
markets with broader societal goals of inclusive growth,
Foster innovation in financial products and structures environmental stewardship, and social equity. Social bonds
to meet the evolving needs of investors and issuers in represent an innovative financial instrument that channels
the social bond market. capital towards projects addressing pressing social challenges
Integrate social bonds into broader environmental, while providing investors with a means to support such
social, and governance (ESG) investing strategies. initiatives while earning returns. Despite of the challenges
By incorporating social bonds into ESG portfolios, faced by social bonds, the continued evolution of social bonds
investors can pursue holistic approaches to sustainable holds promise for driving positive social impact and advancing
investing and align their financial objectives with sustainable finance objectives in the financial market.
broader social and environmental goals.
Provide policy support, incentives, and regulatory References:

frameworks to encourage the issuance and uptake of www.nabard.org
social bonds. This may include tax incentives, subsidies, www.ifc.org
credit enhancements, or regulatory mandates that www.spglobal.com
incentivize issuers to prioritize social impact and
incorporate sustainability considerations into their www.themandarin.com
financing decisions. www.icmagroup.org

Google's busy building LLM for 1,000 languages

Google's chief scientist Jeff Dean has been at the tech giant for over 24 years now and has had a hand in some of our
world's most consequential technologies - like Google's search bar. Most recently, he's been co-leading the Gemini project,
Google's latest family of multi-modal large language models. On a recent trip to India, Jeff said that he cares very deeply
about the ability of these models to work across many languages - as many as a 1,000 - and that Google is actively
working on the project.
"We want to make the models that we're training today and tomorrow to work across the thousand most commonly
spoken languages in the world. And that's spoken languages, not just written, because many languages don't have very
much written material but are still heavily spoken by tens of millions of people."
Such a language model, Jeff said, can have profound implications for India and countries in Africa. "We will be able to
provide access to information in ways that were previously underserved to people who speak all kinds of different languages
in all kinds of different communities. And that's why we're actively working on this across a thousand languages."
On top of that, India's strong army of technical talent means it is well positioned to take advantage of this new digital
era we are all embarking on, which has important implications worldwide. "India has tremendous technical talent; there's
a long tradition of computer science students in India doing amazing things. Anyone with an interest in computer science
is turning their attention to machine learning and what it can do. India is well-positioned in this area because there are
many students that are already in this field and people already in this industry. It's an important thing across the world."
And as important as expensive computational equipment is for this new AI era, Jeff also said that improving the quality
of algorithms can have profound implications for advancing this technology as well - which is crucial for countries like
India that lack advanced AI infrastructure.
"It is definitely the case that training on more data and more compute has helped, but I think it's also a little under-
realised how important algorithmic advances are in improving the capabilities of these models. We've gotten maybe
factors of 10 improvement over the last few years from more powerful computation per wat or per dollar, but we've
probably gotten a factor of 10 as well from algorithmic improvements."

52 | 2024 | MAY | BANKING FINANCE